"Errors are an inevitable part of human nature, and radiology reporting has its share of them," Sangwaiya wrote from the U.K. "In an attempt to create awareness of the incidence and causes of errors in radiology reporting, we conducted this narrowly focused research: to identify the number of discrepancies between the vertebral level of abnormality stated in the body and impressions sections of radiology reports."
Using the keywords "MRI spine" and "addendum," a radiology report search engine (Folio VIEWS 4, NextPage, Draper, UT) performed a Boolean search of 2.1 million radiology reports prepared at Massachusetts General Hospital in Boston, where Sangwaiya worked at the time during a two-year period commencing January 2006. The search engine, which also has the ability to search for precise phrases or expressions, identified and analyzed more than 23,200 reports of MRI exams.
For this quality control assessment, 2,537 MRI spine exam reports with and without addendums were reviewed by researchers. Approximately 1.1% of these reports contained errors, and fewer than 50% of these were identified and corrected with addendums.
Sangwaiya will discuss the process and the details of undertaking narrowly focused analyses for error assessment and quality control evaluations.
